LabWired The easy way to build hardware
Circuit simulator

Simulate circuits with the firmware that controls them.

LabWired connects circuit models to real firmware running on supported microcontroller models. Inspect voltage, GPIO and peripheral behavior before you build the hardware. Start with a browser-based linear circuit, or connect an agent to design and test a supported device.

The browser's linear analog solver is not a full SPICE replacement. Model results are not physical hardware acceptance tests.

What you can simulate

Analog signals

Run supported resistor, capacitor, inductor, source and controlled-switch networks. Probe the circuit and inspect voltage changes over time.

Firmware and peripherals

Run firmware on supported STM32, ESP32 and RP2040 board models. Connect modeled GPIO, ADC, displays and buses to your circuit tests.

One timeline for both

View analog waveforms beside digital edges in the oscilloscope. Measure timing and voltage differences with cursors; export traces as CSV.

PCB and agent workflows

Use LabWired MCP to design and route supported PCBs, run model-based firmware checks, and perform separate KiCad design-rule checks.

Try a circuit simulation

  1. Open the capacitive-touch lab. Arduino Nano firmware drives a 1 MΩ resistor and a touch-pad circuit.
  2. Inspect both signals. The browser oscilloscope displays the D4 drive and D2 pad voltage.
  3. Change the input. Hold the touch pad and compare the response. Read the mixed-signal guide for model limits.

Inspect a reproducible circuit, not just an animation

The Arduino Nano firmware drives D4 and measures the D2 input through a 1 MΩ resistor. Its pinned analog model specifies a 20 pF idle pad and a switched 100 pF finger capacitance. Those capacitances are simulation inputs, not measurements of a physical touch sensor.

In the oscilloscope, CH1 is the D2 pad and CH2 is the D4 drive. Compare the voltage trace, firmware total= readings and D13 LED while the pad is released and while you hold it. The native acceptance script checks separate released and pressed serial count ranges and the LED state. That test is not a browser acceptance result or certification of real-world sensor accuracy.

Observed browser response: released, pressed, released

On 14 November 2026, an unauthenticated Chromium session opened this lab with the oscilloscope running. We held the on-screen touch pad for two seconds, then released it. The oscilloscope's D4 drive (CH2) frequency estimate changed with the modeled capacitance. These are one run's displayed values, not a calibrated sensor benchmark or a statistical repeatability test.

Input stateD2 pad mean (CH1)D4 drive frequency estimate (CH2)
Released, initial2.039 V15.04 kHz
Pressed, after 2 s2.428 V3.61 kHz
Released again, after 2 s1.830 V13.14 kHz

Read the recorded browser observation and method (JSON). The return to the released state is a control against a permanently pressed input. We did not independently assert UART counts or the LED in this browser session; the linked native acceptance test checks those separately. This is not a HeyPCB whole-board simulation.

Want to reproduce the measurement? The STM32 firmware + RC circuit walkthrough shows the PA5-to-ADC1 wiring, 1 ms expected time constant, original source files, UART acceptance assertions, a verified capacitor-change model regression, and a proposed firmware fault injection. It distinguishes the native model test from a browser runtime check and from physical hardware acceptance.

A circuit-to-PCB example

The Cobalt F405 flight-controller case shows a real PCB routing and verification workflow. LabWired ran modeled STM32 firmware against the recovered I²C paths. A separate ngspice DC model checked the battery-sense divider, while KiCad independently checked routing and design rules. These are distinct tests—not one combined proof that the manufactured board will work.

The ESP32-S3 PCB pin-mapping case shows why a parsed MCU is not automatically a runnable circuit.

Importing an existing PCB? See the KiCad PCB import and simulation coverage report. For your own device, request a feasibility review or connect the MCP server.

Circuit simulation limits

The browser analog adapter handles linear circuits. It does not support transistor or diode models, AC or DC sweeps, or vendor SPICE .include libraries. For nonlinear circuits and external SPICE models, the native LabWired co-simulation adapter can use ngspice; that wrapper is not available in the browser.

Use a dedicated SPICE environment when you need full analog-device analysis, and use a physical bench for hardware acceptance. LabWired is most useful when the question is how circuit behavior affects your firmware. See the supported elements and adapter reference and the Falstad CircuitJS comparison for interactive circuit analysis versus firmware-connected tests. For multi-node firmware CI rather than analog circuit analysis, see the Renode virtual-platform comparison.